BRIEF REPORT

Smaller holistic processing of faces associated with facedrawing experience

Guomei Zhou & Zhijie Cheng & Xudong Zhang &

Alan C.-N. Wong

Published online: 4 January 2012# Psychonomic Society, Inc. 2011

Abstract The type of experience involved with an objectcategory has been regarded as one important factor inshaping of the human object recognition system. Laborato-ry training studies have shown that different kinds oflearning experience with the same set of novel objectsresulted in different perceptual and neural changes. Whetherthis applies to natural real-world objects remains to be seen.We compared two groups of observers who had differentlearning experiences with faces, using holistic processing as adependent measure. We found that, while ordinary observershad extensive individuation experience with faces anddisplayed typical holistic face processing, art students whohad acquired additional experience in drawing faces, and thusin attending to parts of a face, showed less holistic processingthan did ordinary observers. These results converge withlaboratory training studies on the role of type of experience inthe development of different perceptual markers for differentobject categories. It is thus insufficient to categorize expertisesimply in terms of object domains (e.g., expertise with faces).Instead, perceptual expertise should be classified in terms ofthe underlying process or task demand.

Keywords Face perception . Face recognition .

Experience . Holistic processing . Perceptual expertise

Object perception is challenging. Apart from the hugevariability in visual input—due to changes in viewingconditions related to lighting, viewpoint, and so forth—awide range of perceptual tasks can also be performed withdifferent objects. Sometimes we categorize objects intogroups, ignoring small differences, while at other times wediscriminate in fine detail to identify individuals; we mayattend to certain parts, to obtain specific information aboutobject function or its affordance (e.g., the ear of a cup forgrasping), or we may attend to all parts in a certainsequence, to portray an object through drawing, sculpting,or other art form. Experience is therefore essential, becauseit allows for the recruitment of optimal processing strategieswhen handling different object perception tasks.

Experience has been shown to modify object representa-tions in the visual system. For example, individuals whohave developed expertise in recognizing objects in a certaincategory (e.g., cars, birds, fingerprints, music notes, letters,and characters) have shown specific perceptual behaviorsand greater activity in specific brain regions, as comparedwith novices (Busey & Vanderkolk, 2005; Gauthier,Skudlarski, Gore, & Anderson, 2000; Gauthier, Wong,Hayward, & Cheung, 2006; A. C.-N. Wong & Gauthier,2007; A. C.-N. Wong, Jobard, James, James, & Gauthier,2009; Y. K. Wong & Gauthier, 2010; Xu, 2005; for adifferent view, see McKone & Robbins, 2007; Robbins &McKone, 2007). Laboratory studies have impressivelydemonstrated that hours of training with novel computer-generated objects can result in similar behavioral and neuralmarkers of expertise (Gauthier & Tarr, 2002; Gauthier,Williams, Tarr, & Tanaka, 1998). While these findings haveunderscored the importance of experience in shaping ourvisual system, recent studies have further showed that theexact manner in which the visual system is modifieddepends on the type of experience. For example, learning

G. Zhou (*) : Z. Cheng :X. ZhangDepartment of Psychology, Sun Yat-sen University,135 Xingang Xi Road,Guangzhou 510275, Chinae-mail: zhougm@mail.sysu.edu.cn

A. C.-N. WongDepartment of Psychology, Chinese University of Hong Kong,Shatin, NT, Hong Konge-mail: alanwong@psy.cuhk.edu.hk

Psychon Bull Rev (2012) 19:157–162DOI 10.3758/s13423-011-0174-x

to separate objects into categories leads to differentbehavioral and neural effects, as compared with learningto identify and name the individuals (Tanaka, Curran, &Sheinberg, 2005; A.C.-N. Wong, Palmeri, & Gauthier,2009; A. C.-N. Wong, Palmeri, Rogers, Gore, & Gauthier,2009). Learning to associate shapes with meanings, ascompared with learning to visually discriminate them, alsoresults in different behavioral and neural effects (Song, Hu,Li, Li, & Liu, 2010). Apart from laboratory training, severalstudies have also shown that perceptual strategies andneural correlates do have an association with real-lifeexpertise for such object categories as cars and birds (e.g.,Gauthier et al., 2000). However, again, these studies haveonly demonstrated that experience does something, withoutaddressing which type of experience brings the specificchanges in question. Similarly, for faces, it is still unknownwhether the type of experience can indeed account for thedifferent perceptual and neural characters associated withreal-life face perceptual experts.

The present study aims to examine the relationshipbetween different types of real-life learning experience andface perception. In our experiment, we tested two groups ofobservers on their face processing. One group was ordinaryobservers, who were experts in identifying individual faces.The other group was art students, who were experts notonly in identifying individual faces, but also in facedrawing. All art students have to pass the college entranceexamination of fine arts, which includes four subjects:sketching, quick sketching, color, and originality. Headpainting is included in the first three subjects. After theyhave passed the entrance examination, they also have totake a course in sketching, which includes head sketching,for at least one semester. An important step of headsketching is depth depiction—that is, to depict the detailsof facial features. Through a large amount of practice, artstudents learn the structure of eyes, nose, mouth, and ears,and become familiar with patterns of facial features indifferent orientations and emotions, and with distinguishingcharacteristics. They learn to quickly and accurately graspthe main facial features of each individual. An intriguingquestion is whether our two groups would differ in theirface perception behavior. Specifically, we asked whetherthe face drawing experts, given their additional experiencein isolating individual face features during drawing, wouldshow a smaller tendency to view faces holistically, ascompared with ordinary observers.

Face perception has been regarded as a special case ofobject perception, with holistic processing as one of themarkers of its special status (Farah, 1991; Farah, Wilson,Drain, & Tanaka, 1998). Specifically, observers tend to payobligatory attention to all parts of a face (Hole, 1994;Richler, Tanaka, Brown, & Gauthier, 2008), and spatialrelationships among parts are also emphasized (Maurer,

LeGrand, & Mondloch, 2002; Mondloch, Maurer, & Ahola,2006). Holistic processing is greater for individuals withbetter face perception performance (Richler, Cheung, &Gauthier, 2011) and is also enhanced after acquiringexpertise in perceiving nonface objects (Gauthier & Tarr,2002; A. C.-N. Wong et al., 2009b). Recent studies,however, have shown the limitations of the effect ofexperience. For example, the effect of expertise inperceiving nonface objects on holistic processing may onlygeneralize to very specific object domains (e.g., modern butnot antique cars, for experts with modern cars; Bukach,Phillips, & Gauthier, 2010; Robbins & McKone, 2007).Besides, in novel training studies, not all kinds ofexperience enhance holistic processing: While experiencewith individuating similar objects can do this, mereexposure or basic-level categorization cannot (Nishimura& Maurer, 2008; A. C.-N. Wong et al., 2009b).

In this experiment, we examined whether all kinds ofexperience would be associated with greater holisticprocessing for faces. Holistic processing, defined here asthe obligatory attention to all parts of an object, wasmeasured because it has been identified as a typicalbehavioral marker of face-like expertise (Gauthier & Tarr,2002; Richler et al., 2008; A. C.-N. Wong et al., 2009b).We adopted a composite paradigm typically used formeasuring holistic face processing (Hole, 1994), with amodified complete design to avoid response biases(Gauthier & Bukach, 2007; Richler et al., 2011; Wenger& Ingvalson, 2002). Participants performed same/differentmatching on the top halves of two sequentially presentedcomposite faces, ignoring the irrelevant bottom parts. Forordinary observers, holistic processing can be found asbetter performance in congruent trials (i.e., both tops andbottoms are the same or different) than in incongruenttrials (same for one part and different for the other). Thiscongruency effect should also be larger when the top andbottom parts are aligned than when they are misaligned.We expected that ordinary observers would show holisticprocessing, typically found in the literature (Richler et al.,2011; Richler et al., 2008). Of interest was whether artschool students, who have acquired additional expertise inface drawing, would show less holistic processing, due totheir experience in attending to individual parts andfeatures for detailed drawing.

An important strength of the present study was that itallowed us to probe whether lab training studies of theeffect of type of experience on object perception apply toreal-world expertise. However, this also means that thestudy was correlational in nature, because experience wasnot manipulated. Smaller holistic processing associatedwith art students would thus only suggest, not necessarilyimply, that face drawing experience causes a decrease inholistic processing.

158 Psychon Bull Rev (2012) 19:157–162

Method

Participant

A group of 48 students from Sun Yat-sen University (24male, 24 female; 20–24 years of age) and 50 students fromGuangzhou Academy of Fine Arts (33 male, 17 female;19–24 years of age) participated for monetary compensa-tion. The art students had 2–16 years of experience indrawing faces. All participants had normal or corrected-to-normal vision and signed a consent form before theexperiment.

Design

A 2 (group: control or art student) × 2 (alignment: alignedor misaligned) × 2 (congruency: congruent or incongruent)mixed design was used, with group as a between-subjectsfactor and alignment and congruency as within-subjectsfactors.

Materials

Eight grayscale Chinese male faces with a neutral expres-sion were split into top and bottom halves in order to formcomposite faces. For composites with aligned top andbottom parts, each composite subtended a visual angle of5.4° × 8.2° at a viewing distance of 50 cm. Each misalignedcomposite subtended an angle of 8.2° × 8.2°.

Procedure

The experiment was run by E-Prime 1.2 software (www.pstnet.com/eprime.cfm). Each participant completed a facecomposite task. A face identification task was alsoadministered to assess their face recognition ability, inorder to verify that the two groups had comparable levels offace recognition ability.

In each trial of the face composite task, a 200-ms centralfixation cross appeared, followed by a 250-ms blank screen.Then the first composite was displayed for 200 ms. After a500-ms interstimulus interval (ISI), the second stimulus wasdisplayed for 2,000 ms or until response. Participants wererequired to determine, as quickly and accurately aspossible, whether the tops of the two composites were thesame or different. Half of the trials were same trials, andhalf were different trials. In congruent trials, the tops andbottoms were both the same or different; in incongruenttrials, the tops were the same and the bottoms different, orvice versa. There were 32 trials for each of the fourconditions (alignment × congruency), resulting in a total of128 trials. Participants completed 8 practice trials withfeedback before the actual experiment.

In the face identification task, which was modeled on theCambridge Face Memory Test (Duchaine & Nakayama,2006), the stimuli were the original faces used in thecomposite task. All faces subtended an angle of 5.4° × 8.2°.Each trial began with a 200-ms fixation cross, followedwith a 250-ms blank screen. Then the target face appearedfor 200 ms, followed by a 500-ms blank screen and thesecond stimulus array, with four faces displayed untilresponse. Participants had to determine whether the targetface was present in the four-face array as accurately aspossible by pressing 1, 2, 3, 4, or 5 (absent). In target-present trials, the target was displayed in the four positionswith equal probabilities. A total of 64 target-absent and 64target-present trials were presented. Before the actual test,8 practice trials were presented with feedback.

Results

Composite task

A total of 13 participants’ data (5 control and 8 art students)were discarded because of low accuracy (< 56%). For responsetimes (RTs), only correct trials with an RTwithin three standarddeviations from the mean of each group were included in theanalysis. The mean sensitivity (d' = ZHit – ZFA) and correctRTs are shown in Fig. 1 and Table 1, respectively. Typically,holistic processing is indicated by more accurate and fasterresponses for congruent than for incongruent trials, especiallyfor the aligned condition (Richler et al., 2011; Richler et al.,2008; A. C.-N. Wong et al., 2009b). As has typically beenfound in previous studies, we observed such a result patternin both groups. However, the art students showed smallereffects in general.

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

Alig

ned

Mis

alig

ned

Alig

ned

Mis

alig

ned

Control Art students

d'

Congruent Incongruent

Fig. 1 Mean sensitivity (d') and standard errors for control studentsand art students as a function of alignment and congruency

Psychon Bull Rev (2012) 19:157–162 159

This observation was confirmed by a 2 (group) × 2(alignment) × 2 (congruency) mixed ANOVA conductedon d'. The three-way interaction was significant, F(1, 83) =4.40, p < .05. Separate analyses for the two groups revealedthat the interaction between alignment and congruency wassignificant for the control participants, F(1, 42) = 22.298,p < .001, but only marginally significant for the art students,F(1, 41) = 3.618, p = .064. We also directly compared theholistic processing of the two groups, defined as the effect ofalignment on the congruency effect [(aligned incongruentRT – aligned congruent RT) – (misaligned incongruent RT –misaligned congruent RT)]. This measure indicates how muchthe interference from the irrelevant part depends on the intactconfiguration of parts (Bukach et al., 2010; Cheung, Richler,Palmeri, & Gauthier, 2008; Richler, Bukach, & Gauthier,2009; Richler et al., 2011; A.C.-N. Wong et al., 2009b). Thismeasure was significantly smaller for art students (0.43) thanfor the control group (1.12), F(1, 83) = 4.40, p < .05.

The 2 (group) × 2 (alignment) × 2 (congruency) mixedANOVA conducted on RTs revealed a significant maineffect of congruency, F(1, 83) = 5.88, p < .05. Other maineffects and interactions were not significant (ps > .12).

Identification task

The art students and the control group displayed similarface identification ability. A one-way ANOVA on accuracyin the identification task revealed no significant differencebetween the two groups (art students, M = .73, SD = .14;control, M = .68, SD = .12), F(1, 83) = 2.449, p = .121.

Correlation analysis

In order to examine the association between drawing experi-ence and the magnitude of holistic processing for the artstudents, we calculated the Pearson correlation between yearsof drawing and the size of the holistic effect on d'. There weretwo missing values, because 2 art students did not reporttheir years of drawing experience. Zero years of drawingexperience was assumed for the control group. A significantnegative correlation was found between years of drawingexperience and holistic processing for all participants, r

(83) = −.224, p = .042; the more drawing experience, theless holistic processing. This correlation, however, wasmainlydriven by the discrete group differences, as can be seen in thescatterplot (Fig. 2) as well as in the nonsignificant correla-tions within the art students [r(40) = −.05, p = .76].

An interesting set of analyses, related to a recent debate(Konar, Bennett, & Sekuler, 2010; Richler et al., 2011), is thecomputation of correlations between face identificationperformance and holistic face processing. We did notobserve such a correlation in d', r(85) = .017, p = .877, norin RTs, r(85) = .033, p = .764. Separate analyses for eachgroup showed similar results (d', ps > .617; RT, ps > .248).Therefore, although we adopted a complete design similar tothat of Richler et al. (2011), our results were similar to thoseof Konar et al. (2010), who used a partial design.

Discussion

Recent years have seen a number of empirical studiesdetailing how the object recognition system is shaped by

Fig. 2 Scatterplot showing the correlation between years of drawingexperience and the magnitude of holistic processing for both groups

Table 1 Mean correct RTsand standard errors for controlstudents and art students asa function of alignment andcongruency

Aligned Misaligned

Congruent Incongruent Congruent Incongruent

Control

Mean 710.48 730.52 720.30 721.48

SE 20.24 21.52 20.54 21.67

Art Students

Mean 745.15 759.69 737.70 746.22

SE 20.48 21.77 20.78 21.93

160 Psychon Bull Rev (2012) 19:157–162

nature (Polk, Park, Smith, & Park, 2007; Zhu, Song, Hu,Li, Tian, Zhen, & Liu, 2010) and nurture (Gauthier et al.,2000; Xu, 2005). On the nurture side, laboratory trainingstudies have indicated that how our mind handles an objectcategory depends not only on the degree, but also on thetype of experience involved (Song et al., 2010; A.C.-N.Wong et al., 2009b; A. C.-N. Wong et al., 2009c). Thepresent study extended previous findings to faces andshowed that different types of experience with the samereal-world category are associated with different perceptualphenomena.

We found holistic face processing for all observers. Thisis not surprising to those who believe that holisticprocessing is a design of nature for perception of such anevolutionarily important category, or to those who attributeholistic processing to extensive experience in the finediscrimination of faces. What is intriguing is the reducedholistic processing, but identical face identification ability,found for art students experienced in face drawing ascompared with ordinary observers. The results suggest thatthe art students were more flexible in shifting attentionamong parts of a face, and thus were more successful infollowing the instruction of part matching in the facecomposite task. Because face perception entails differentjudgments (concerning identity, gender, expression, inten-tion, etc.) involving different sets of diagnostic information(Joyce, Schyns, Gosselin, Cottrell, & Rossion, 2006;Schyns, Bonnar, & Gosselin, 2002), it is possible that theadditional experience of drawing would enhance one’sattentional flexibility to better utilize different facialfeatures for different tasks, making one a more versatileface processor.

A limitation of the present study is its correlationalnature. There was no pretest to determine whether artstudents had shown holistic processing identical to that ofthe control students before training. Alternative explana-tions are therefore possible with regard to the associationbetween art students and reduced holistic processing. Forexample, it is possible that those who gravitate to art areparticularly good at decomposing an object or scene into itsparts. Thus, it may not be that art training causes changes inholistic processing. Further studies that directly manipulatetraining will be needed to address this issue. It is possiblethat a relatively short training regimen (months, or evenweeks) may suffice, given the similar levels of holisticprocessing for art students with different years of training,as shown in our correlation analyses.

An implication of the present results concerns therelationship between holistic processing and efficient faceidentification. A close tie between these abilities has beenemphasized by previous findings of a correlation betweenholistic face processing and face identification performance(Richler et al., 2011) and increased holistic processing of

novel objects as a result of individuation training (A.C.-N.Wong et al., 2009b). We found, however, that the artstudents showed less holistic processing than did theordinary observers, yet had similar face identificationperformance. Also, we did not observe any correlationbetween the magnitude of holistic processing and faceidentification performance for either the art students or theordinary observers, contrary to what Richler et al. (2011)found. This prompts the question of whether holisticprocessing, or obligatory attention to all features, isnecessary for efficient face identification. A possibility isthat attention to all features does help one identify faces,but individuals differ in their experience of fulfilling othertask demands (e.g., drawing, photography, verbal descrip-tion, etc.) with faces. These other experiences might affectthe degree of obligatory attention to all features but not faceidentification ability, weakening the correlation between thetwo.

To conclude, different levels of holistic processing werefound for observers having different types of experiencewith faces. An intriguing possibility is that face drawingexperience, as well as experience acquired later in life afterone has had extensive experience identifying individualfaces, can still alter one’s style in face perception. Theexperience of learning to fulfill different task demandsassociated with different objects might therefore be essen-tial for continuously optimizing object perception abilities.Future training studies would be required in order todirectly test the causal relationship between types oftraining and face perception. One may, however, start toconsider categorizing perceptual expertise in terms of thetask demands involved (e.g., face identification expertise,face drawing expertise) instead of merely in terms of thedomains (e.g., face expertise).

Author Note This research was supported by the National NaturalScience Foundation of China (31170978) and the 11th Five-year Plan inPhilosophy and Social Sciences of Guangdong Province (08SXQ001) toG. Z., and by grants from the Chinese University of Hong Kong (DirectGrant 2020939) and the Research Grants Council of HongKong (GeneralResearch Fund 452209) to A.W.

References

Bukach, C. M., Phillips, W. S., & Gauthier, I. (2010). Limits ofgeneralization between categories and implications for theories ofcategory specificity. Attention, Perception, & Psychophysics, 72,1865–1874. doi:10.3758/APP.72.7.1865

Busey, T. A., & Vanderkolk, J. R. (2005). Behavioral and electro-physiological evidence for configural processing in fingerprintexperts. Vision Research, 45, 431–448.

Cheung, O. S., Richler, J. J., Palmeri, T. J., & Gauthier, I. (2008).Revisiting the role of spatial frequencies in the holistic processing offaces. Journal of Experimental Psychology. Human Perception andPerformance, 34, 1327–1336.

Psychon Bull Rev (2012) 19:157–162 161

Duchaine, B., & Nakayama, K. (2006). The Cambridge Face MemoryTest: Results for neurologically intact individuals and aninvestigation of its validity using inverted face stimuli andprosopagnosic participants. Neuropsychologia, 44, 576–585.

Farah, M. J. (1991). Patterns of co-occurrence among the associativeagnosias: Implications for visual object representation. CognitiveNeuropsychology, 8, 1–19. doi:10.1080/02643299108253364

Farah, M. J., Wilson, K.-D., Drain, M., & Tanaka, J. N. (1998). Whatis “special” about face perception? Psychological Review, 105,482–498. doi:10.1037/0033-295X.105.3.482

Gauthier, I., & Bukach, C. (2007). Should we reject the expertisehypothesis? Cognition, 103, 322–330. doi:10.1016/j.cognition.2006.05.003

Gauthier, I., Skudlarski, P., Gore, J. C., & Anderson, A. W. (2000).Expertise for cars and birds recruits brain areas involved in facerecognition. Nature Neuroscience, 3, 191–197.

Gauthier, I., & Tarr, M. J. (2002). Unraveling mechanisms for expertobject recognition: Bridging brain activity and behavior. Journalof Experimental Psychology. Human Perception and Perfor-mance, 28, 431–446.

Gauthier, I., Williams, P., Tarr, M. J., & Tanaka, J. (1998). Training“Greeble” experts: A framework for studying expert objectrecognition processes. Vision Research, 38, 2401–2428.

Gauthier, I., Wong, A. C.-N., Hayward, W. G., & Cheung, O. S.-C.(2006). Font-tuning associated with expertise in letter perception.Perception, 35, 541–559.

Hole, G. J. (1994). Configurational factors in the perception ofunfamiliar faces. Perception, 23, 65–74.

Joyce, C. A., Schyns, P. G., Gosselin, F., Cottrell, G. W., & Rossion,B. (2006). Early selection of diagnostic facial information in thehuman visual cortex. Vision Research, 46, 800–813.

Konar, Y., Bennett, P. J., & Sekuler, A. B. (2010). Holistic processingis not correlated with face-identification accuracy. PsychologicalScience, 21, 38–43.

Maurer, D., LeGrand, R., & Mondloch, C. J. (2002). The many faces ofconfigural processing. Trends in Cognitive Sciences, 6, 255–260.

McKone, E., & Robbins, R. (2007). The evidence rejects theexperience hypothesis: Reply to Gauthier & Bukach. Cognition,103, 331–336. doi:10.1016/j.cognition.2006.05.014

Mondloch, C. J., Maurer, D., & Ahola, S. (2006). Becoming a faceexpert. Psychological Science, 17, 930–934.

Nishimura, M., & Maurer, D. (2008). The effect of categorisation onsensitivity to second-order relations in novel objects. Perception,37, 584–601.

Polk, T. A., Park, J., Smith, M. R., & Park, D. C. (2007). Natureversus nurture in ventral visual cortex: A functional magneticresonance imaging study of twins. Journal of Neuroscience, 27,13921–13925.

Richler, J. J., Bukach, C. M., & Gauthier, I. (2009). Contextinfluences holistic processing of nonface objects in the compositetask. Attention, Perception, & Psychophysics, 71, 530–540.doi:10.3758/APP.71.3.530

Richler, J. J., Cheung, O. S., & Gauthier, I. (2011). Holisticprocessing predicts face recognition. Psychological Science,22, 472–477.

Richler, J. J., Tanaka, J. W., Brown, D. D., & Gauthier, I. (2008). Whydoes selective attention fail in face processing? Journal ofExperimental Psychology: Learning, Memory, and Cognition,34, 1356–1368.

Robbins, R., & McKone, E. (2007). No face-like processing forobjects-of-expertise in three behavioural tasks. Cognition, 103,34–79. doi:10.1016/j.cognition.2006.02.008

Schyns, P. G., Bonnar, L., & Gosselin, F. (2002). Show me thefeatures! Understanding recognition from the use of visualinformation. Psychological Science, 13, 402–409.

Song, Y., Hu, S., Li, X., Li, W., & Liu, J. (2010). The role of top-down task context in learning to perceive objects. Journal ofNeuroscience, 30, 9869–9876.

Tanaka, J. W., Curran, T., & Sheinberg, D. L. (2005). The training andtransfer of real-world perceptual expertise. Psychological Science,16, 145–151.

Wenger, M. J., & Ingvalson, E. M. (2002). A decisional component ofholistic encoding. Journal of Experimental Psychology: Learning,Memory, and Cognition, 28, 872–892. doi:10.1037/0278-7393.28.5.872

Wong, A. C.-N., & Gauthier, I. (2007). An analysis of letter expertisein a levels-of-categorization framework. Visual Cognition, 15,854–879. doi:10.1080/13506280600948350

Wong, A. C.-N., Jobard, G., James, K. H., James, T. W., & Gauthier, I.(2009a). Expertise with characters in alphabetic and non-alphabetic writing systems engage overlapping occipito-temporal areas. Cognitive Neuropsychology, 26, 111–127.

Wong, A. C.-N., Palmeri, T. J., & Gauthier, I. (2009b). Conditions forface-like expertise with objects: Becoming a Ziggerin expert—but which type? Psychological Science, 20, 1108–1117.

Wong, A. C.-N., Palmeri, T. J., Rogers, B. P., Gore, J. C., & Gauthier,I. (2009c). Beyond shape: How you learn about objects affectshow they are represented in visual cortex. PLoS ONE, 4, e8405.

Wong, Y. K., & Gauthier, I. (2010). A multimodal neural networkrecruited by expertise with musical notation. Journal of Cogni-tive Neuroscience, 22, 695–713.

Xu, Y. (2005). Revisiting the role of the fusiform face area in visualexpertise. Cerebral Cortex, 15, 1234–1242.

Zhu, Q., Song, Y., Hu, S., Li, X., Tian, M., Zhen, Z., et al. (2010).Heritability of the specific cognitive ability of face perception.Current Biology, 20, 137–142. doi:10.1016/j.cub.2009.11.067

162 Psychon Bull Rev (2012) 19:157–162